Analysing performance of SLEUTH model calibration using brute force and genetic algorithm–based methods

Abstract

Present study is aimed to compare the performance of SLEUTH model from two different calibration methods, that is, brute force and GA in term of computational efficiency of calibration processes, capturing urban growth, a form of growth or growth pattern and its spatial distribution. SLEUTH has been parameterized for Ajmer city (India) and its performance has been compared in term of eight parameters/methods, that is, computational efficiency, model fitness that is, OSM, urban shape index, best fit coefficient values, hit-miss-false alarm method, kappa statistics, accuracy percentage and visual analysis. GA-based calibration has been found to be computationally more efficient and relatively better in capturing urban growth and form of growth as compared to brute force. Brute force calibration seems to be slightly better considering urban hits as compared to GA, however, GA is better with respect to lesser false alarms.

Keywords:

• Urban growth
• SLEUTH
• cellular automata
• brute force
• genetic algorithm

Note

¹http://www.ncgia.ucsb.edu/projects/gig/

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

We are highly indebted to the Ministry of Human Resources and Development (MHRD), India for providing financial assistantship. Also, we acknowledge the FIST program of DST, Government of India for funding research laboratory.